Don’t get left behind: AI playbooks from leading fleet operators

Hopefully, you can all hear us. We’re very, very excited to be here with you all today.

Just a couple of things to get out of the way before we get started. I do wanna mention that this session will be recorded, and you’ll receive, an email with the recording afterwards. So please look out for that.

And, yeah, I’m really, like, very excited to get into this discussion with you all, talk about how AI is impacting fleet operations and and the physical economy at large. So with that, I’d love to do some intros for a couple speak for for our speakers today.

Jamie, you wanna start?

Sure. Hi, everyone. My name is Jamie Bergstrom. I’m the manager of the Department of Transportation for Step Energy Services, the Canadian operations. So I oversee transportation and fleet operations across Western Canada and support field operations in a demanding and fast paced oilfield environment.

Great. Thanks, Jamie. Brendan?

And I am Brendan Wiggins. I am the vice president of fleet and equipment here at Congrux.

So I own fleet, the entire operation here for the fleet, and we have a national operation here in the US. So across, you know, I think about thirty states and several thousand pieces of equipment, and we do fiber optic construction. So putting fiber in the ground, everything from the ISPs all the way to your house.

Awesome. And, I’m Michael. I lead the AI team at Motive, and that involves all the things touching AI in our products. I’ve been working on AI applications for a very long time, many ten ten plus years. And looking forward to chatting with you all more about Motive and and what is going on out there with AI and physical economy. So before we get into the discussion, just want a quick overview of Motive for you all.

You know what we do and and how it all fits together. Motive is really like an AI powered integrated operations platform for the physical economy. We help the companies that keep the physical world running operate more safely, productively, and profitably.

And our platform, it brings together a number of different components under one umbrella. There’s the driver safety.

There’s the fleet management, the equipment monitoring, spend management, workforce management, maintenance, and AI vision, which is our non vehicle type of applications outside the cab.

And the the thing that connects all of it is is the AI at the heart of it that turns data from vehicles and equipment and works worksites into decisions that your teams can act on.

And my team is responsible for the AI behind all of these products. So I get a close look at what’s working in the field, what’s coming next, and and what’s on the minds of of the folks who are using it. So what I wanna dig into today, if we go to the next slide, is how AI is moving from something that people just talk about or use like ad hoc to something that teams are running their operations on, something that they count on and depend on.

And why does it matter? It’s because we believe, and and I think you all are probably seeing this, that the cost to waiting is real. And if you wait and you don’t get on this train before too long, you’ll find that you’ve lost down on huge opportunities and huge opportunities to improve efficiency and gain revenue. So in our research before our vision conference this past May, we estimated that around forty six billion dollars in financial loss is tied to problems that AI can already solve, like preventable downtime, unproductive unproductive idling, and compliance violations.

And our industry research found that teams lose about sixty three percent of the work week to manual tasks and top operations win back up to twenty five hours per week once they can leverage AI to take over some of that repetitive work.

And thirty eight percent of operations leaders named advances in AI and automation among the biggest opportunities for growth.

So I don’t wanna spend time on the scary version of this story, though. What I really wanna do is talk about what happens when real teams actually lean into AI and use it to foundationally across their operation to transform how they do their work. And that’s why I’m so excited to have Jamie and Brendan here with me today to share, like, practical experience of how this ends up happening in the field. So let’s get into it.

I’m gonna start off with a question we’ll pose to the group here, and then we can we’ll take it from there. We’ll see where things go. But the first question I wanna know from Jamie and Brendan is where do you see AI showing up most clearly in your day to day operations today, and how has that changed the way your teams work?

Jamie, you wanna start?

Yeah. I’ll start. So it’s kinda I had this, like, moment when this whole AI started, and I am a paper person. So I was very, like, you know, up against the wall on this whole AI thing and and technology.

But it was when I realized that, paper was lagging in safety and data, and I realized that good or bad, it’s to protect the fleet and our professionals. That’s what the AI is here to help us with now. And, so, yeah, just more real time visibility, across logistics. And, the dual facing cameras have just helped scale safety oversight so much and just give us a clearer view of what’s happening in the field.

So we’re using AI to improve reporting on driver behavior and consumption, not just to review incidents after the fact.

It’s nice to have that real time data right there and not have to, not have to wait for the information. Right? And one of the biggest shifts is being able to confirm if safety protocols are actually being followed or not, good or bad.

Yeah. It’s hard it’s hard with paper to keep track of all of it, maybe.

So Still like the paper as the backup. But How about you, Brendan?

What are you seeing, on your in the day to day world?

Yeah. I think day to day, you know, is is I don’t wanna say obvious, but just the the time savings on things like summarizing documents and and just it’s a force multiplier for my team. Right? So we’re able to get a lot more work done more quickly with the same people, and, and and, really, one of the places that’s come into play the most, I would say, is the safety stack, right, is the camera and the safety program where we went from having effectively nothing, right, up up to that point and then going to a dual facing camera.

And and just trying to scale a program up like that would have been very difficult for us to go manually grind through, making sure we had the right people in the right chairs to go process all them that information in in near real time versus, you know, it’s allowing it to distill so much information for us down into really actionable events and have allowed us to bring that program up to speed a lot more quickly than we would have been able to in the past. So, you know, it’s about it’s about giving the existing team a leg up right now versus really handling anything on its own, but but really using that to get the right people doing the right work.

Yeah. Yeah. Definitely. You know, what what I’m observing is that the key to unlocking the power of AI is is data.

And a lot of organizations today, what I see is that they have a lot of data. They’re actually data rich, but they don’t have a lot of time. So I say they’re time poor, data rich, time poor. And the the key is to give them the tools and the capability to unlock the value of that data so that the busy operators can use it without having to dig through dashboards or reports or paper paper trails if as as they’re doing their jobs.

So I really think that we can unlock that level of productivity by reducing the manual work, getting the time back, and, of course, making things safer so that people come home at the end of the day and there’s less accidents on the road. So, that’s great great to hear from you both. I wanna change gears now and talk kind of about the other side of this, which is, you know, the financial, aspect of it. And I think one of the things that I believe very strongly is that these applications need to have value proposition for you, for the customers, for the people who use them.

And so my question for you is, you know, where are you seeing the biggest returns on AI today in terms of monetary returns, costs, or time or other utilization metrics? What do you think? Brendan, you wanna Sure. Any thought on that?

I I think, again, a lot of it comes down to right now is time is money. Right? And and I have skilled laborers, specifically, the mechanics are a great example, right, where I mean, effectively, minute that those guys aren’t turning wrenches, they’re being underutilized.

And so how can I how can I, you know, take their time and make sure that, again, the right people doing the right work and minimize some of the admin work that they’re doing? So, you know, there’s that and some of my administrators, my shop, you know, managers and those folks where they’re taking invoices and we’re able to process those invoices into the maintenance system a lot more quickly by just snapping a photo of it and letting the AI transform that invoice from dozens of different vendors into our single format, you know, there’s a huge amount of time savings that comes from that. And

and that alone doesn’t save the money. What what does save the money is then having that skilled person go do what they’re actually good at. Right? What are they actually here to perform?

And so, you know, it’s a it it that savings comes out on the ability of my team to accomplish more in the same amount of time. So that’s a lot of where we’re going right now, and that goes all the way up to me. Right, where, you know, I was actually just recently, that, you know, we had a new leader come into the organization in the seed suite, and they came and said, hey. I need to know x y z about our our fuel spend and fuel volume over the course of the last year, especially the fuel price volatility and everything we’re dealing with.

And were there reports that I could have gone and run and put all this together? Absolutely, there were. Right? But what I was able to do was go into the Motive AI and go, hey.

I need to go look at my fuel card data over the last two years and summarize it in a way x y z that will show me how much of it is due to fuel price volatility versus how much is growth. Right?

Because if I’m running more miles and and and running more business, fuel price increasing is actually a good thing versus if it’s all fuel price fluctuation, then that’s not a great thing, right, for our for our so how do we go quantify how much was coming from each bucket?

And the Motive AI put that together for me in a few minutes in what would have taken me probably an hour or two to put together from running a few different reports that were already accessible to me. But then I took that time and went and invested it somewhere else.

Super interesting. Yeah. It’s great to hear about how these tools can actually give you back some of that precious time, that resource that we’re all trying to find more of all the time. So how about you, Jamie? What are you seeing in terms of, like, efficiency gains, cost savings, other things?

Yeah. So, again, like, the team can start coaching sooner instead of waiting for lagging reports, especially in incidents and stuff like that, like anything on the road, so we can just get on top of things quicker. But exactly like Brendan was saying, like, reporting off fuel use, the driver behavior, safety trends is it’s just it’s helped so much. And it’s it’s kinda funny.

So we’ve been with motor for five years now, and STEP’s been around for fifteen. And I kinda think back now, like, the amount of reports and data that we can send financing and the maintenance team and just our operations now, I was like, how did we do this before? You know? Like, it’s just so easy now because the data is just there.

And I can’t even, like, I can’t even remember how we used to gather it before or how long it took us or, you know, even how accurate it was.

And now it’s just we can see how much more accurate it is, the time like, it’s less time consuming and everything, and we can just give our, our operations, our finance teams, everything just so much more data to help save the company time and money too now.

One thing that’s huge is for our mechanics, like, mechanical data.

It’s just you know, motive tracks the engine hours. It tracks the mileage. It tracks, what would you call it? Like, critical high, low fault codes and stuff. Right?

So instead of, again, the mechanics waiting for stuff, we can go in there and run all these reports and get them the information, and they can get on top of things quicker so that our trucks are continuously moving. Right?

They’re less lag time, which is awesome. So, again, yeah, it’s, going in and asking AI to help us with these reports. And I just feel like every day we’re finding new ways to run reports and using the AI is now helped us so much because we’re just we’re learning more and more too with now the whole AI options.

Yeah. Yeah. Definitely. I think, the opportunity for automating a lot of that and gathering that information for you is is where we see a lot of customers digging into this and and finding value.

You know, I I think there’s no substitute for human judgment at the end of the day, so we need to be mindful of that. Right? And so I’m sure you’re still reviewing and spending time on these things, but, hopefully, the big lift of, like, the prep work and all of that pulling the data, can be offloaded. That is what I would expect. Is that right?

Oh, yeah. Completely. Like I said, yeah, we still have our our paper trace that comes in or the reviews after, but to actually just get that data quicker and to be able to start reviewing it sooner is what helps.

Yeah. Yeah. Yeah. Definitely. Yeah.

Just you know, I I would say one of our bigger risks is management trusting it too much too quickly. Right? Yeah. And so, you know, the the role that our subject matter experts play is still critical in in how this all works. Right? We we have to have both and not an either or. Yeah.

Yeah. Yeah. Totally. Yeah. I can see that some folks in the crowd agree.

So I wanted to shift gears a little bit from, you know, obviously, the financial upside and the opportunities there are are very good and impactful. But I think the safety is where it gets real very quickly. Safety AI.

Right? And so my question for you all is, you know, how do you see AI helping to improve safety and coaching, and how do you make sure that people trust it and that they don’t get kind of fatigued by the AI coaching or flagging them for the wrong things. What do you think, Jamie? Any thoughts?

Yeah. So, again, it’s everyone thinks camera’s in your face. You know? It’s a how do you say it?

It’s a gotcha tool. Right? Like, it’s we’re the people are spying on you, but it’s not. Like, I don’t have all the hours in the day to sit back and watch cameras all day.

So, we’ve gotta we’ve gotta be positive about the cameras, and it’s it’s your friend. It’s like your co driver. Right?

For us, what I love about it all is that you can the company can set their own settings, and they can put their company standards in there. So and, there’s, like, fifteen unsafe behaviors that can be monitored by the AI. Examples are what are some of them?

Just like the fatigue, the seat belt, the distracted driving. Like, there’s so there’s fifteen of them in total. And it’s just about like, the drivers, they get alerted first. So they have real time self correction that can happen.

So they get alerted first. They get alerted I think it’s up to, like, three times before a manager is even alerted. So it’s it’s not just the camera caught you doing something, and you’re getting told on. Right? It’s there to help you. It’s like your co driver. It’s there to keep you safe.

And, like I said, the your team can tune alerts to the company standards, and we have found that there has been like, it can filter false positives by, like, thirty percent within the AI. And, I don’t know. I think it just it helps the newer drivers learn faster, and it helps long term drivers correct habits they never even realized they had because they’ve been driving for so long when none of this technology was ever around. Right?

So, yeah, just it builds a stronger safety trust with your professionals, and it’s really, important to also back the positive driving behaviors also and not just concentrate on things that AI can be pointing out to us now.

Yeah. Totally. Great points. Yeah. So, Brendan, I’m wondering, like, how do you think about trust with these tools, and how do you how do you talk about that with your team?

I I think a lot of it is how you design your program in your fleet, right, where how you talk about the program especially early is is how you build that trust. It doesn’t just happen overnight. There’s not the right words you’re gonna say that are gonna cause the driver court. It just trusted out of the gate.

You you have to build the program the right way and communicate it the right way. And and really that builds over time once you start to see some key exoneration events. Right? You have the whole driver court turn on you overnight when you get a couple really big exonerations and and really focusing on those early is where you start to get the the pull of the cameras into the cab versus them you have to push them into the cab.

And and so, you know, we kinda went through that early where the drivers were pushing back on it. And now, yeah, you still have your holdouts. But for the most part, people understand that the tool is a positive thing. And and, really, the focus of leadership is on protecting them and protecting the organization rather than really looking for something.

Like like Jamie said, I’m not nobody has time to spy on them in the cab. Right? We’re we’re really there to identify where where we have risk and and how do we go lean in on that. So, you know, I think that, you know, honestly, it’s really just about how you communicate it and then, you know, how you how you implement the program long term to build that trust over time.

And and things you’re gonna have false positives really no matter how you deal how what technology you use. Right? So the AI is, again, another tool for us to help tune out the noise. It’ll it’ll get better over time, which is one of the nice things about it.

But, you know, things like speed limits and and, you know, the g force thresholds for certain events that it’s gonna detect, you know, we we really end up with false positives on that and have for ages. So it’s it’s really just about how do we slowly chip away and improve in that over time, especially on things like like speed limit that are a bit of a moving target.

Yeah. Yeah. Speed limit is an interesting one because it’s kind of like, there’s there’s a lot of nuance there, right, and what the other cars are doing and what what the latest speed signs are. But, you know, I I wanna point out that, like, the the accuracy is so important to us at Motive because the last thing we want is somebody to get annoyed by our product and turn off potentially a lifesaving feature.

So we we really look very closely at the accuracy, especially of the in cab event yeah. Alert coaching and in cab interventions. Those are where we need the precision to be very high, the accuracy to be very high. So we build that trust over time, like you said.

And then once people come to trust, I do see countless examples where these things can be lifesaving in in various circumstances. So, yeah, I I think that’s a really important point and something we definitely don’t wanna lose sight of.

I I wanna kind of, you know, I I think talk about the other side of this, which we’ve we’ve touched a tad, but, you know, there’s more to say, which is that even though AI is great and has done amazing things for you guys and many others in the industry, there’s still gaps. Right? There are real gaps. And I was wondering if maybe you guys could shed some light on where do you see not maybe living up to the promise yet, you know, entirely, and where do you see opportunity for the next, like, wave of AI and innovation and enhancements? Brendan, any thoughts on that?

I I think it’s similar to any of the analytics tools we’ve been using over the last decade plus, Michael. Right? It’s there’s garbage in, garbage out. But if you if you give it bad data, it’s gonna give you bad information back.

And and so the data quality still matters. Like, how you code your work order still matters. How you’re doing your DVIR still matters because that’s the information that you’re using to help you make these decisions. So you you you don’t just get a pass on data integrity, when you go and implement AI because that’s what it’s gonna use to help you.

And so it can help clean up that data, but but, ultimately, it’s making a judgment call just like you would. So so the quality of your inputs really still matters, and and you need to make sure that all those foundational pieces are in place, that all your systems are connected and integrated well to to operate seamlessly.

And, you know, it’s it’s not a silver bullet that’s just gonna solve all of your data problems overnight. Right? It it really requires you to still do the work up front to set that foundation for you to be able to to leverage it in the future.

Yeah. Yeah. Definitely. Jamie, what do you think? Where do you see the not quite living up to the hype yet?

Yeah. Like, I I totally agree with what Brendan just said. And, exactly, you need to learn the platform’s abilities to get it to give you the best data, but, exactly, there still has to be the data review, right, of everything that it can give you.

Like I said, it’s just AI is a new technology, and it’s changing all the time, and it’s growing. And, like, your team’s constantly working on this, Michael, and stuff. And and, sure, there’s a few things that, like, for I think if it helped the driver more, like, more driver input, just things like, more direct support for the drivers, like, better suggestion.

But, I mean, we’re all we’re all learning it. Right?

But, yeah, it’s if if we don’t use it, if we don’t give it the right data, it’s not gonna help give us the right data.

Yeah.

So it’s it’s all just a learning process. I don’t think it’s hard to, like, say exactly where it falls short right now because we’re all still trying to learn it. We’re all still growing with it. It’s growing. So yeah. Yeah. That’s a tough one for me.

Yeah. I mean, I think, like, we talked about this. Right? There’s there’s still the need for a healthy amount a a good amount of human judgment in This whole process. Right? And, I think AI is very powerful, and it’s becoming more powerful every day, as you mentioned.

But the one thing that no AI can do, and this is where, you know, it all kind of comes down to the line, is it can’t take responsibility. And so at the end of the day, we have to take responsibility for the outputs it generates and the way that it, you know, the way we use those outputs. And so I do think that in physical operations, you know, with the stakes being much higher than, say, just like a chat where you’re asking what to have for dinner or where to go out with your friends, I do think that the the peep people need to do more to check and double check the outputs that we get and own the decisions.

It’s not to say there’s not a ton of value. And like I said, I think, like, heavy lifting, that’s the way I view it is that a lot of the heavy lifting can be offloaded, but the judgment and decision making, that still has to be, you know, a human element to it. So I totally agree with you guys on that. I wanna ask a question, which is kind of a touchy subject, but we’re all friends, so I know I can bring it up with you guys.

You know, I think, like, a lot of people in our audience, a lot of people hear AI and the, you know, kind of alarm bells, like job replacement, headcount reduction, We’re, you know, we’re not gonna need as many people to do the things we’re doing.

And I know that we’re all thinking about optimizing and making things more efficient, but I’m curious if you guys could talk a little bit about what’s actually happening today in your teams on the ground and what is not happening, you know, in regards to this. So I don’t know. Jamie, you wanna give us your thoughts?

AI and the headcount? No. We we still need train trained professionals. That’s not gonna go away. For the oil fields, like, that’s my background.

AI can’t drive down a lease road or rig up a frac crew. Right? We still need the bodies. We still need the headcount for that. We need the professionals.

But what AI can do is, help get our professionals there safer.

So like I said, it’s just like, it’s your co driver. Right? It’s just it’s there to help you.

That’s what this whole system is about. It’s not taking over your position. It’s not taking over my position in the office. Someone still has to go through that data.

Someone still has to review it. We have to stay on top of it. We’re here to help our professionals with their logbooks, with everything. Right?

So, yeah, AI doesn’t scare me for my job quite yet.

That’s good to hear, Jamie. I yeah. I mean, like, we’re definitely not in a world where those things you described can be offloaded to the AI. So I I see that your your team is probably not feeling this impact at all yet. But, Brendan, how about your side? What do you see?

I kinda feel the same. It is you know, we already had more work than time. And so, you know, it it’s really just allowed my team to be more effective right now. And and so the you know, the I’m I’m nowhere close to where it’s gonna impact head count for me.

I’m not gonna say that it’s never gonna happen long term in in certain parts of the business, but but, realistically, right now, there’s no future where we look like we’re going to be where we’re gonna be reducing headcount. Really, all it’s done is allowed my team to be more effective in the near term. So, you know, again, taking the skilled labor that I have, like Jamie was talking about, I have skilled fleet managers and skilled mechanics and and really just taking some of the things that they were doing that weren’t value added. They weren’t driving value to the bottom line and and collapsing those a little bit and giving them more time back in their day to go do other things.

So, yeah, it’s it’s obviously something I think everyone’s keeping an eye on, but as of right now, I don’t see any horizon where where that’s gonna impact headcount for me.

Yeah. Yeah. I mean, I think, like, one of the things I saw recently that stuck with me is that if you look at how we do work today versus how we did it ten years ago. Right?

Like, if you think about that and you think about how we use spreadsheets and, you know, document editing and collaborative tools and all the web stuff that we have, you would think, oh, well, like, we should have made things so much more efficient. Everyone should have so much free time compared to where they were ten years ago. I don’t think any of us feels like we have more time free than we did ten years ago right now. So somehow, we find ways to refill the bucket.

You know? When when things are opening up, there’s so much more we can do. There’s so much more value we can create. And that’s where I think the promise comes from here is that we really need to find ways to redeploy and and utilize those free the free freedom that AI is providing.

So, yeah, really appreciate those those thoughts. Yeah.

And that’s like what I said earlier. I I can’t even remember how we were doing this ten years ago. How are we getting this data? How I I don’t even think we were compared to how we are now.

Right? And it’s Yeah. Yeah. It’s just filling a different bucket. It’s not really sure. It’s saving some time, but it’s yeah.

Exactly. Filling a different bucket of what we can look at and do now with our information that can be provided to us now.

Right. Right. Yeah. Totally. Totally. I wanna ask another question that is kind of AI adjacent, I guess, or related, which is, you know, we hear a lot about people facing fragmentation in their business, you know, different business units, contracts, legacy systems.

My question is, do you feel like there’s something more we need to do to unlock the power of AI when things are so fragmented? And, what are your thoughts on that? Maybe, Brendan, you have some takes on take on that.

Yeah. So I, you know, I I am experiencing that. We have some legacy business units, and we’re still working through consolidating everyone, into one sheet of music from a from a systems perspective, right, the back end systems. And so at least in right now, where AI is helping the most is planning those integrations. Right? Before, it used to be a huge effort to go create the documentation that we would need to hand off to the technical teams to help us build those integrations.

Whereas now, you know, it it’s much more quick to iterate on on those things. So we can go say, hey. Here’s what we need to integrate and why and and what we want to accomplish, and it can kinda help us generate that documentation for the technical teams a whole lot quicker and and scope question or, like, scoping documents and, common pitfalls and things that we didn’t think of, right, edge cases that we need to take into account. You know, those used to take iterations of passing those documents back and forth between various teams, and it’s kinda collapsed that process for us much quicker.

And so we’re able to be more agile and and get a lot more done in a in a shorter period of time just because me as the business team are providing IT a lot more valuable and actionable information upfront rather than them having to come back to us and ask us questions that we know the answer to. We’re just not used to having to provide, you know, a technical document to to that team. And so, you know, the art of the possible has become easier for us.

Yeah. Yeah. Totally. How about you, Jamie? Any thoughts on the fragmentation, how that affects the business and the value here?

Yeah. So same as Brendan was saying that every every business, every fleet needs something different with the data, with the information. Right? And, yeah, it used to take us a bit and have to share and everything.

And now we can we can basically just kinda click on the settings that gives each team exactly what they need. And then at the end of the day, go in and run that report or allow them to see that information, and it’s just there. And it’s already separated. It’s what they need.

If multiple teams need that information, they can have it. We’re not exactly, like Brendan said, we’re not passing it back and forth. It’s all there for them. Right?

And Motive gives us that, opportunity with their settings to set each department up differently to give them what they need. And, of course, AI works best when teams are clear on their standards. And but, again, you have to use the system.

So the more you use it, it learns from us and then understands our standards and what we’re looking for. And so to help give us the data and perform better with the false positives in that, you need to use the system so that it the AI can help understand what exactly you need.

Yeah. No. I I think the oh, yeah. Sorry.

Go ahead.

Oh, and, yeah, just one added thought that came to me while Jamie was talking, Michael, was, you know, the, you know, the fragmentation can be also at least I mean, in our business, I’m sure everyone else is where I have people who need fleet data who are not fleet users. Right? They’re not gonna be in Motive or or my maintenance system all the time. And so they’re asking questions to my team that they could really self serve if they knew where to fish. And So you’ll have a stakeholder who’s gonna log in to Motive a few times a year.

But, you know, telling them, oh, go run this report and go run that report. It can be challenging for them just because it’s muscle memory. They’re not used to where to go and find some of these things. And that’s somewhere I found lately in my story earlier was that it makes some of those reports a lot more accessible just using the Motive AI because they could just go in there and go, hey. What’s the what’s the site with the worst MPG in the last six weeks? Right? And it’ll go run that for them and deliver that data to them.

So it can be particularly valuable in breaking down those barriers to other functional areas that aren’t your day to day users that you normally have in the system.

Yeah. I mean, one thing I will say is that I have observed myself that the AI AI in general is way more valuable and powerful when it has access to more context, when it can pull more information and has a view that’s holistic and unified.

If you narrow that or limit it, then you don’t get the power that you’re looking for as much. And then you have to be the one to hop between these different contexts or pieces.

With Motive and the Atlas system that we’ve built, it’s all kind of under the hood that it’s powering everything, all of your data in one place, accessible to one AI. But I think to Jamie’s point, not everyone has the same needs and use cases. So it’s both about giving that broad view and perspective to the system, but also giving the ability to tailor the usage and the the reporting and the specific settings to, you know, a business unit or a specific set of drivers. So that’s where I think, you know, the the two things come together and really create a ton of value. So, yeah, I really appreciate it.

I like how you use that example, Brendan, and, you know, not everyone’s a fleet user. Not everyone logs in all the time and and exactly that. Sometimes you hear back from someone, and I’ll be like, how did you get that information? Like, you don’t have those settings. How did you do that?

And that’s exactly it. They go in. They use the AI, and I’m like, oh, okay.

Although, ask my team. Yeah. Who gave this person this information?

Yeah.

Yeah. I think self serve self serve the ability to self serve and get things, you know, done yourself has never been, like, easier. It’s never been easier than to do that, but that comes with some pitfalls too. I think, Jamie, like, what what you’re saying that sometimes people don’t always understand what they’re looking at.

And so we have to, you know, constantly look at what is coming in and and triage and and understand it. Help them understand it. So really appreciate it. Those are really interesting discussion dialogue here about those topics.

I wanna now switch gears and and see if our audience wants to post any questions or or anything that we can answer that’s on their minds.

I think we’re gonna just take them from a queue or maybe someone will, pull pull them up. But if if people have questions, please post them, and I think we’ll get we’ll get fed into our queue here. Oh, there’s a q and I’m sorry. This is not as familiar with this tool. There’s a q and a tab. Okay.

So alright. Let’s start with, okay. I I think this is a good question from Steven York here. We’ll start with this one. So, Brendan, you said at the beginning that using AI has helped keep your mechanics turning wrenches.

How specifically has the AI, and motive helped with paperwork and running of operations as from a fleet manager perspective?

Yeah. So it’s it’s a great question and, you know, it it seems to be evolving quickly, but what you know, and not all of this full disclosure is is specifically the mode of AI compared to what I’m talking about AI holistically in our organization. So, you know, I have a tool that is allowing folks to scan invoices into the maintenance system, right, and and add work orders into the maintenance system using, you know, machine vision, which has been a huge time saver over time. But when you think about mechanics turning wrenches, it’s also helping mechanics prioritize things.

Right? So when we look at the fault codes coming through from Motive, that’s coming into my back end system, and then we’re triaging those. Right? We’re so we’re leveraging AI to help us triage faults.

I’m getting fault severity and then likely root causes of those as well. So I can say on this vehicle platform, I’m seeing this fault or faults, and it’s giving a a a general output of what it thinks might be wrong with that vehicle.

It’s not a hundred percent right all the time, but I would say the vast majority of the time, it helps narrow the field, when you get into specific failures on specific make models or known issues on those make models. It can save a technician a ton of time from running down the diagnostic tree versus, you know, if you see a twenty twenty three Ford Transit with this x y and z fault code, it’s this, you know, then they’ll go check that first. And a lot of times confirm that that is actually the failure mode, and then we’re immediately into the repair versus the diagnostic phase where we might have just gone by the book down the existing diagnostic tree. So, you know, it’s not perfect, but there have been plenty of instances where just that simple, hey. It’s this make model year and and spec of the vehicle, and here are the faults we see. And, of course, MOTI’s providing the inputs for that, and then the output being, hey. Check this first.

We’ve found a couple instances where that saved us a huge amount of time, but I really think that that’s gonna continue to evolve over time as, again, the AI gets access to more tools and we get that better integrated. Right now, you know, we’re still having the guys kind of bounce around between a couple different tools to to accomplish that. And, you know, in some of our younger techs who are a little more early adopters and willing to go do some of that stuff versus the guys who still wanna listen to it with a screwdriver.

So Right.

You know, we’ll get there, but but it it already is kinda paying dividends on on just saving time on the diagnostic tree.

Diagnostics. Yeah. Predictive maintenance, those kinds of applications. Yeah. Definitely.

Can I just add something to that? Oh. Sorry.

Of course. Yes.

And the potential to utilize the AI and the ECM data in our trucks, it can help us with better fuel efficiency upgrades that we can use in our fleet. And so that’s something great for, our maintenance teams and and that to, be able to see that data to help with our like, with better fuel efficiency and stuff.

Got it. Got it. Okay. Yes.

And, also, one last thing, Michael, and I’m just gonna use the opportunity for my soapbox, but it’s a pet like, become a sticking point for me of that we all in the industry, we talk about predictive maintenance.

Right? And I was guilty of helping coin that term, I think, from my prior life. But if if you’re making decisions based off of a fault code, you’re not predicting anything. Right?

Like, that failure is already too late.

Or is it progress? Right? So when we think about predictive maintenance, we’re looking at known failures in under certain conditions.

Like, you were Right.

It’s a bit more advanced when we talk about predictive maintenance. And I view the fault code triage as as a means to an end there. Right? It’s along that path, and we’re getting pretty good at it. But the the that’s something I think that’s very accessible for everybody today is is triaging engine and vehicle faults much better. And then when we talk about predictive true predictive maintenance, it’s just a little bit harder than it sounds.

That’s the holy grail. Right? That’s the holy grail.

But we’ll get there eventually. But but, yeah, the the fault code triage stuff, I think, is has potentially add a lot of value for everybody.

Yeah. Definitely. Yeah. So, I wanna there there’s a question in here from Joe Castro, asking where does Motive stand with the many competitors competitors out there?

With AI being relatively new, most of the equipment and AI training will soon be obsolete. How often will equipment need to be replaced? So there’s a couple things in there I wanna unpack and I wanna kinda, like, ask you guys a couple questions. But, like, I will say that, you know, as it pertains to Motive’s AI, we believe we have world class, best in class capabilities, and we invite people to to test that out for themselves.

But as far as the equipment and the the obsolete obsoletion or the kind of, like, nature of AI to change over time, there’s absolutely going to be advances. And in order to unlock some of that, there will be needs for new hardware, but the hardware cycles are generally, you know, several years, multiple years before we can refresh or change anything. So a lot of what we do is to upgrade software, which gets pushed over the air to all of our hardware devices. But I am curious from you guys like Jamie and Brendan, how often do you see that you have to replace equipment related to AI like dashcams or gateways or anything like that?

What do you see in the field? Jamie, you wanna weigh in?

Well, with our, like, with our new three in one ELD dashcams with the AI technology, I guess we’ve been using it since January. I don’t think there’s really been any replacement yet, so that’s a hard one to answer, or maybe an easy one.

I get you know, from your experience over many years, you know, how often are you having to go and replace these things or upgrade them and and that kind of stuff?

Yeah. I mean, our basic upgrade was we just had the ELDs, and then we moved to the, like, outward facing cameras. Right? And now we’re getting into the three in ones.

So So every few years, maybe you’re seeing kind of a transition.

Is that right?

Yeah. As technology grows. Right?

So And that’s because you wanna unlock these new capabilities, is it?

Is that what it like, to bring them online?

Oh, yeah. Completely. Yeah. Because the the information, the data, the opportunity it gives us, why wouldn’t we?

Yeah. Brandon, what do you think about that?

I view it as, I guess, you know, new technology and fleet has always been a bit of a double edged sword. Right? Like, you just you you you wait, and that can be safer, but then you leave opportunity on the table. It’s always a business case of of risk versus reward on adopting new technology, and that’s not new.

We’ve been doing it since alternative fuels and, you know, running a CNG program in twenty fifteen wasn’t all sunshine and rainbows. Right? But but the you know, there there are things that, you know, the experience you gain as an organization on adopting new technology usually pays dividends long term on your ability to evaluate that technology. And the the AI dash cameras are are just in the newest version of that.

Right? And so when I install hardware in the vehicles, I plan on that piece of hardware being able to live the life cycle of that vehicle in my fleet. And so the perfect world would be I’m gonna install it when I buy the vehicle, and it’s gonna live in that vehicle until I sell it. And There are times where that’s not the case.

Right? There are times where the technology has moved enough that I have a a good enough business case to go replace it. But that’s what it is. It’s a business case.

Right?

I have to look Oh.

What’s the investment for me to go outlay and go and upgrade that hardware and the and the time spent to do that and everything else that goes into that. But I’m only gonna do that when there’s a a true benefit to me ripping replacing that hardware. And and it’s Yeah. You know, it’s it’s a it’s a decent hurdle to to cover, but I think that any of the hardware I’ve used, so far has been capable of performing its role in the business through the life cycle of the unit. And then I’m making a judgment call on whether we wanna adopt any new hardware.

So it’s like, you know, you’re viewing it as, hey. You know, I’m I’m gonna get this value proposition. I’m gonna put this device in my vehicle or in my my equipment. And then if another device comes out, I’ll evaluate the incremental value I can get. And if it’s worth the cost, I’ll do it. Is that the framework? Oh, yeah.

Okay. Yep.

Okay. I wanna there’s there’s another question in here I I think is interesting. I wanna, take on here from Corey Sherlock, which is, does AI correct itself if a driving event is dismissed due to no violation? For example, AI flags a stop sign being ran. However, the sign was for an adjacent road. Absolutely. The answer to that is yes.

The way we train our models and the way we build our systems is to continuously learn, and the idea is that nobody should have to do anything. But as this data is coming in and being reviewed and audited and by the way, we have people auditing things internally, and then you all as customers will give us feedback, and then we’ll use that in our system as well. So all those audits that come in and all of those corrections, they get fed back into what’s called our training dataset, and then a new model is trained and built on top of that that incorporates these learnings and and will get better over time. So I definitely believe, and we have seen that, you know, the feedback we gather from customers, the feedback we gather from our annotators improves model performance dramatically. But I’m curious from you guys, Jamie and Brendan, like, do you see things changing over time? Do you see that as the technology gets rolled out, it improves over time in your applications? What do you think?

Jamie, do want With the, like, with the new AI, dash camps and that, I love that it so we’d always get things from our drivers. Like, this is an eighty zone, not a fifty zone anymore, things like that. Right? And now with the new AI dash cams, it it picks up the the signs, like, the speed limits, and it will adjust, like, our whole fleet.

So then once one camera picks it up, it changes everything in the Motive system and which has been wonderful because it’s way less phone calls. Because we’d always just say to the guy, yeah. Okay. So the construction zone’s gone now.

Like, we just gotta wait till things you know, till this new sign’s put up or until things change. You know? And so I love that that part of the new technology. And, I mean, a stop sign’s a stop sign.

Right? We actually just had a case where we had a stop sign on one of our gates at our shop, but when it’s open, the camera could still see the stop sign. But, really, the stop sign was for when it was closed. Right?

Just to be aware that the the sign was there. And so, yeah, we actually contacted Motive. We gave, like, a an example of what it was, and the problems were fixed because it was constant. Our trucks are going through that gate, and the guys were like they were getting mad that it was getting caught, that they were catching the stop sign.

And we’re like, okay. It’s not a problem. We can fix this. You know? And that’s exactly what it was.

So it was there for a safety purpose that the AI was helping us, and we were easily able to fix it.

Yeah. Anything you wanna add, Brendan?

My examples of if that’s what you were asking.

Perfect. Yeah. That’s that’s exactly it. Yeah.

I think it’s a great example, Jamie. We’re right. Like, I think back in the early days of the ADAS systems, right, where we used to come out of the one parking lot, and there was a set of train tracks right there, and the metal would reflect it back at the ADAS system, and it would slam on the brakes on the drivers every time. And it took a little bit of work to get that to work properly.

But I think it’s all about the quality of your partner, right, in in in this case, Motive. And in your case, right, you email them and go, hey. I have this weird edge case that’s causing a lot of pain and suffering for my team, and, know, how do we address that? And the responsiveness of your partner is really where you’re gonna succeed or fail there when you’re adopting new technology.

So, you know, in my experience, anytime I have an edge case or something like that that needs addressed, Motive is all over it, and, and it’s been a great experience. So I think that’s the only thing I’d probably add beyond what Jamie said.

Yeah. Actually, I do see a question in here about the speed sign issue specifically.

So someone’s asking, can we expect the camera to read speed sign speed limit signs anytime soon? And the answer is yes. So, Jamie, you have that in your system today enabled. Yeah?

Yeah.

So this is something that, we’ve recently launched and who Julia Williams, I believe, is asking. If you wanna, more information, we can we’ll connect with you afterwards. Alright?

I guess, you know, I think one last question we have time for here, and, I think, like, the kind of overarching thing I see, there’s a bunch of questions related to, is really how how accurate is this stuff? Like, do you find it to be relatively accurate? Are you happy with it?

Any other thoughts on that as a closing kind of wrap up here?

I say, yes. I found it to be generally accurate. Now like any new technology, you need to take it with a healthy dose of skepticism. Right?

You can’t just take it as as, you know, verbatim. There are going to be edge cases. There are gonna be exceptions. But, honestly, there were when we had people doing it.

And on top of that, there was a lot of things that we didn’t even have the bandwidth to look at before this. So I’m in the camp that the net benefit is there. Right? That we are we’re net net in a better place than we were even with the false positives that we still get, and that addressing them has gotten easier than it was before in the past as well.

So it’s really about how the program is designed. You don’t wanna go create penalties for one incident. You know, usually risk isn’t driven from one incident anyways. It’s driven from a pattern of behavior over time, and really one false positive on a driver shouldn’t change that.

So I think just it goes back to how you’re implementing the program, but overall, my opinion is, yes. It’s accurate enough to be actionable, and it adds value to the organization.

Yeah. Okay.

And Anything there?

Yeah. Love what you said there, Brendan. And, again, I’m gonna go back to what I said at the very beginning.

Once I realized good or bad so good or bad being, oh my gosh. Why did why did AI just send that as a a high what does it even come across as?

Like, a like, a high priority thing. Right?

And so we can we can just change that in a click of, okay. That was coachable. We coached the driver. Dismissed that.

No. That’s not what happened in this in the incident or whatever. Right? And so that’s the good or the bad, but it goes back to, in the end, I’m glad it caught it because it’s there to protect our fleet and to protect our professionals.

And so I’m glad it caught it so that we could actually see, and we can help train the AI or we can or thank goodness the AI was there to help us, and now we can train our drivers. We can protect our fleet. You know?

That’s just what I like.

To realize the good or the bad, it’s there to protect the fleet and the professionals, and we can work around things. We can work with it.

So that’s all I can say.

That’s great. Thank you, Jamie. I think we we can wrap up now. I wanna thank you both for taking the time. Thank you to the audience members. You all asked some really great questions.

Really interesting discussion about how we can leverage AI, how it’s actually being used in physical operations today, and also some of the limitations and drawbacks that we heard about. So it’s a real talk here. I I really I I couldn’t be happier with Jamie and Brendan as my co colleagues here, and looking forward to continuing the conversation with you all. Okay?

Thank you so much for joining. We’ll see you soon. Bye bye.

Thank you, everyone.